Web-based tool for detecting bias in reviews
View Patent ↗One embodiment provides a computer system for detecting associations between a reviewer and an entity under review. During operation, the system estimates a relationship strength between the reviewer and the entity under review, and determines whether the relationship strength between the reviewer and the entity under review exceeds a predetermined threshold.
1. A computer system for detecting associations between a reviewer and an entity under review, comprising:
a processor;
a memory;
an identifying mechanism configured to identify a first set of terms associated with the reviewer, wherein identifying the first set of terms involves issuing a search query that includes the reviewer's name, and extracting a predetermined number of terms from the search results;
an estimation mechanism configured to estimate, for a review conducted within a user-community, a first relationship strength between the entity under review and a respective term of the first set of terms associated with the reviewer, wherein estimating the relationship strength involves:
issuing a first search query to determine a first number of documents external to the user-community that indicate the entity under review and the respective term associated with the reviewer;
issuing a second search query to determine a second number of documents that indicate the entity under review; and
calculating a ratio of the first number of documents to the second number of documents;
a determination mechanism configured to determine whether the first relationship strength exceeds a predetermined threshold; and
a bias-reporting mechanism configured to report a potential bias of the reviewer toward the entity under review in response to the first relationship strength exceeding the predetermined threshold.
2. The computer system of claim 1 , wherein the entity under review is a service, a product, or an entity associated with the service or product.
3. The computer system of claim 1 , wherein the estimation mechanism is further configured to:
issue a third search query for the entity under review;
retrieve a number of returned hits for the third search query; and
calculate a ratio of the number of returned hits for the first search query to the number of returned hits for the third search query.
4. The computer system of claim 1 , wherein the search query is a Web search engine query.
5. The computer system of claim 1 , further comprising an identifying mechanism configured to identify a second set of terms associated with the entity under review.
6. The computer system of claim 5 , wherein the estimation mechanism is further configured to:
estimate relationship strengths between each term of the second set of terms and the reviewer.
7. The computer system of claim 1 ,
wherein when the first relationship strength between the reviewer and the entity under review is below the threshold,
the estimation mechanism is further configured to estimate a second relationship strength between the reviewer and a second entity and a third relationship strength between the second entity and the entity under review;
the determination mechanism is further configured to determine whether the second relationship strength between the reviewer and the second entity and the third relationship strength between the second entity and the entity under review both exceed the predetermined threshold; and
the bias-reporting mechanism is configured to report a potential bias of the reviewer toward the entity under review based on the determination.
8. A computer-executable method for detecting associations between a reviewer and an entity under review, comprising:
identifying a first set of terms correlated with the reviewer, wherein identifying the first set of terms involves issuing a search query that includes the reviewer's name, and extracting a predetermined number of terms from the search results;
estimating, for a review conducted within a user-community, a first relationship strength between the entity under review and a respective term of the first set of terms associated with the reviewer, wherein estimating the relationship strength involves:
issuing a first search query to determine a first number of documents external to the user-community that indicate the entity under review and the respective term associated with the reviewer;
issuing a second search query to determine a second number of documents that indicate the entity under review; and
calculating a ratio of the first number of documents to the second number of documents;
determining, by a computer, whether the first relationship strength exceeds a predetermined threshold; and
responsive to the first relationship strength exceeding the predetermined threshold, reporting a potential bias of the reviewer toward the entity under review.
9. The method of claim 8 , wherein the entity under review is a service, a product, or an entity associated with the service or product.
10. The method of claim 8 , wherein estimating the relationship strength further comprises:
issuing a third search query for the entity under review;
retrieving a number of returned hits for the third search query; and
calculating a ratio of the number of returned hits for the first search query to the number of returned hits for the third search query.
11. The method of claim 8 , wherein the search query is a Web search engine query.
12. The method of claim 8 , further comprising identifying a second set of terms correlated with the entity under review.
13. The method of claim 12 , further comprising:
estimating relationship strengths between each term of the second set of terms and the reviewer.
14. The method of claim 12 , further comprising:
when the relationship strength between the reviewer and the entity under review is below the threshold,
estimating a second relationship strength between the reviewer and a second entity, and a third relationship strength between the second entity and the entity under review;
determining whether the second relationship strength between the reviewer and the second entity and the third relationship strength between the second entity and the entity under review both exceed the predetermined threshold; and
reporting a potential bias of the reviewer toward the entity under review based on the determination.
15. A non-transitory computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method for detecting associations between a reviewer and an entity under review, the method comprising:
identifying a first set of terms correlated with the reviewer, wherein identifying the first set of terms involves issuing a search query that includes the reviewer's name, and extracting a predetermined number of terms from the search results;
estimating, for a review conducted within a user-community, a first relationship strength between the entity under review and a respective term of the first set of terms correlated with the reviewer, wherein estimating the relationship strength involves:
issuing a first search query to determine a first number of documents external to the user-community that indicate the entity under review and the respective term associated with the reviewer;
issuing a second search query to determine a second number of documents that indicate the entity under review; and
calculating a ratio of the first number of documents to the second number of documents;
determining whether the first relationship strength exceeds a predetermined threshold; and
responsive to the first relationship strength exceeding the predetermined threshold, reporting a potential bias of the reviewer toward the entity under review.
16. The computer-readable storage medium of claim 15 , wherein the method further comprises:
when the first relationship strength between the reviewer and the entity under review is below the threshold,
estimating a second relationship strength between the reviewer and a second entity and a third relationship strength between the second entity and the entity under review;
determining whether the second relationship strength between the reviewer and the second entity and the third relationship strength between the second entity and the entity under review both exceed the predetermined threshold; and
reporting a potential bias of the reviewer toward the entity under review based on the determination.
17. The computer-readable storage medium of claim 15 , wherein the entity under review is a service, a product, or an entity associated with the service or product.
18. The computer-readable storage medium of claim 15 , wherein estimating the relationship strength further comprises:
issuing a third search query for the entity under review;
retrieving a number of returned hits for the third search query; and
calculating a ratio of the number of returned hits for the first search query to the number of returned hits for the third search query.
19. The computer-readable storage medium of claim 15 , wherein the method further comprises identifying a second set of terms correlated with the entity under review.
20. The computer-readable storage medium of claim 19 , wherein the method further comprises:
estimating relationship strengths between each term of the second set of terms and the reviewer.